You set the strategy.
Claude Code handles execution.

Turn any high-level goal into an AI-planned job queue that runs autonomously: self-correcting, fully local, on your existing Claude Code subscription.

Requires Claude Code subscription. macOS 12+ (Apple Silicon & Intel)

OpenHelm
AI Assistant
Ask anything...

Claude Code stops when you do.

Close the terminal and AI leverage ends. No background scheduling.

Goals don't plan themselves.

Turning a high-level objective into scheduled, dependency-aware jobs takes real work.

Other tools add AI costs on top.

You already pay for Claude Code. Extra AI billing is the wrong model.

How it works

From goal to running job in four simple steps.

1

Set a goal

Describe what you want in plain English. The AI breaks it into a concrete job plan. Or build jobs manually. You stay in control.

AI Assistant
Run weekly SEO audit on my SaaS
I'll create a goal with 4 recurring jobs to audit your site every Monday at 6am...
Ask anything...
2

Review the plan

See the proposed mix of one-off and recurring jobs. Approve as-is, tweak individual jobs, or edit cron expressions directly.

Weekly SEO Audit
Ensure robust coverage of SEO health...
ActivePause
Jobs (4) + Add Job
Crawl sitemapMon 6amSucceeded
Analyze page speedMon 6amNever
Check meta tagsMon 6amNever
Generate reportMon 6amNever
3

Jobs run in background

Close your laptop. Jobs run on schedule via your Claude Code subscription, entirely on your machine.

Code Review
Daily · 12:20amEnabled
Run History (3)
Succeeded1h ago19m 41s
Failed1h ago<1s
Failed11h ago<1s
4

Failures self-correct

Failed jobs are analyzed and retried automatically. Persistent failures surface with full context so you can step in.

Security Scan
Auto-retried
Detected failure · analyzed · retried automatically
Failedrun 3fa1c02
↳ Permission denied on /tmp, retrying with corrected scope...
Succeeded↻ auto-retryrun 6ac7b87

Built for Claude Code users

Goal → job plan, automatically

Describe a high-level objective and the AI structures it into concrete one-off and recurring jobs. Review, tweak, and approve. Or skip the AI and build jobs manually — you stay in control.

Goals
Test Coverage
Run test suitesonnet
Identify gapssonnet
Write testssonnet
Gen reportsonnet
Test Coverage
Keep test coverage above 80% through automated weekly checks.
ActivePause
Jobs (4) + Add Job
NameScheduleLast Run
Run test suiteMon 6amSucceeded
Identify gapsMon 6amNever
Write testsMon 6amNever
Generate reportMon 6amNever

Self-correcting execution

Failed jobs are analyzed, adjusted, and retried automatically. Only what the system can't resolve escalates to you — with full context.

Run History
FailedAttempt 1Test runner timeout
Auto-retry: extended timeout to 120s
SucceededAttempt 23m 24s

Memory that compounds

Run history and project context accumulate over time. Every job runs smarter than the last.

Code Review
Run 6ac7b87a
Succeeded19m 41s
Summary
Completed comprehensive code review. Found 16 issues: 3 critical bugs (path traversal, state leak), 5 security vulnerabilities, and 8 code quality improvements. All 5,862 tests passing.
1563 linesRaw
[Tool: Bash] [Tool: Read] 1→# Project Memory 2→3→## MCP Server Config 4→- Linkup Search replaced as default web search provider 5→- Global API key: fbe29462-04dd-4865...

No extra AI costs

Jobs run through your existing Claude Code subscription. No separate billing.

Fully local

Your goals, plans, and results never leave your machine. No cloud, no data exposure.

Fair Source

Source-available under BSL 1.1. Read every line, fork it, and extend it freely.

Stay in control

Skip the AI planner entirely. Create and schedule jobs manually when you prefer.

Fair Source. Fully local. Nothing leaves your machine.

Source-available under BSL 1.1. Read every line, contribute improvements, and fork it for non-commercial use. Free for individuals, students, and startups under $1M revenue. Your goals, plans, and results stay on your machine. No cloud processing, no data exposure.

Fair Source (BSL 1.1)
Runs locally
View on GitHub

How we compare

The only Claude Code scheduler that uses your existing subscription. No extra AI costs layered on top.

FeatureOpenHelmOpenClawDevinrunCLAUDErun
Uses Claude Code subscription
Goal-to-task planning
Self-correcting failures
Memory system
Open source

Comparison based on publicly available information as of March 2026.

Put your Claude Code subscription to work. 24/7.

Goal planning, autonomous scheduling, and self-correcting execution, on top of Claude Code. Free for individuals and small teams, entirely on your machine.

Requires macOS 12+ and a Claude Code subscription.